Skip to content

A Python port of the Apache Lucene ASCII Folding Filter that converts alphabetic, numeric, and symbolic Unicode characters which are not in the first 127 ASCII characters (the ‘Basic Latin’ Unicode block) into ASCII equivalents, if they exist.

License

sandinmyjoints/fold_to_ascii

Repository files navigation

fold_to_ascii

A Python port of the Apache Lucene ASCII Folding Filter that converts alphabetic, numeric, and symbolic Unicode characters which are not in the first 127 ASCII characters (the "Basic Latin" Unicode block) into ASCII equivalents, if they exist.

Usage

> from fold_to_ascii import fold
> s = u'Astroturf® paté'
> fold(s)
u'Astroturf pate'
> fold(s, u'?')
u'Astroturf? pate'

Differences from JS fold-to-ascii

This library always removes astral characters, even when a replacement character is specified. PRs welcome if you want to fix this.

Differences from unidecode

There are a few Python unidecode libraries out there (1, 2). They are based on a Perl program that makes some interesting choices about what to replace, for example, £ (POUND SIGN) is replaced with the string PS. Also, they do not allow specifying a replacement character to use other than the empty string for unmapped characters.

Development environment

$ virtualenv -p python3 ~/.local/venvs/fold_to_ascii
$ source ~/.local/venvs/fold_to_ascii/bin/activate
$ pip install -r requirements.txt

Test

$ make lint && make test

Release

Requires a pypi account.

  1. Bump the version in setup.py.
  2. Build and upload:
$ rm dist/* && python setup.py sdist bdist_wheel
$ twine upload dist/*

About

A Python port of the Apache Lucene ASCII Folding Filter that converts alphabetic, numeric, and symbolic Unicode characters which are not in the first 127 ASCII characters (the ‘Basic Latin’ Unicode block) into ASCII equivalents, if they exist.

Resources

License

Stars

Watchers

Forks

Packages

No packages published